This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. TDP-43 is a protein that is normally involved in splicing if RNA and exon rearrangement in neural cells. In disease, it is responsible for the pathophysiology of amyotrophic lateral sclerosis (Lou Gehrigs disease), frontotemporal lobar degeneration (the most common form of dementia in persons under 60 years of age), and Alzheimer disease. In the latter disease, TDP-43 folds abnormally and aggregates in the cytoplasm of neural cells in the brain, causing misfunctioning of the nerve pathways. The protein consists of two primary regions: a dimeric RNA recognition motif (RRM) and a glycine-rich C-terminal region. The glycine-rich region is the site of a number of mutants, each causing a different variation of disease. Little is known about the three-dimensional structure of the glycine-rich region, but we do have structural data on the RRMs. We are interested in characterizing the glycine-rich region to better understand the structural differences that cause the mutations to encode different disease pathologies. On the RRM, we will perform virtual screening to identify small molecules that may serve as drugs to counteract the function of TDP-43 in disease. Work described in this abstract will involve three experimental fronts. 1) Virtual screening of the RRMs of TDP-43 with about a million small molecules to identify potential drug candidates;2) molecular dynamics unfolding experiments on the glycine-rich region and the RRMs to identify the folding pathways that lead to misfolding in Alzheimer disease;and 3) molecular dynamics simulations on the wild type glycine rich moiety as well as on mutant forms of this protein segment. For the virtual screening experiments we will use the program Surflex-Dock. On a Pentium-4 system, it requires about 7 hours to dock 25,000 small molecules into the mapped active site of the RRM. We plan to use 10,000 CPU hours to dock small molecules into the RRM structure. The highest scoring candidates will be collected and chemically synthesized for testing in the native TDP-43. Protein folding experiments with molecules of any size require extremely large amounts of computer time. In practice, it is typical to assume that unfolding of a protein follows the same pathways as does folding, only in reverse. Since one begins with a known structure, the experiments require much less time. To conduct unfolding experiments on the glycine-rich region, we will begin with a structure that was deduced using the Tripos package Fugue. This package essentially uses a BLAST algorithm and an extensive database to compute likely structures of the molecule. The unfolding experiments will involve molecular dynamics simulations on the model using the AMBER-10 package which will be run for 100-250 nanoseconds to fully sample the molecule in space. A perturbation (increasing temperature) will be used to start the unfolding cycle. Targeted molecular dynamics will be used to drive the structure to the form found for Alzheimer disease. We plan to use 50,000 CPU hours to conduct unfolding experiments. Lastly, we will employ molecular dynamics simulations using AMBER-10 to simulate the 140 amino acid wild type form of the glycine-rich moiety, deduce its conformational structure, and compare it to dynamics simulations of mutant glycine-rich proteins. These simulations will show the conformational structures of the amino acid side chains and will be used to compare wild type with mutants. Nothing is known about these conformations, so these experiments will be extremely important. We will use 10,000 CPU hours to simulate these systems. John M. Beale, Ph.D. Associate Professor of Medicinal Chemistry Saint Louis College of Pharmacy